Page MenuHomePhabricator

Build a maintenance script to clean up duplicate data.
Closed, ResolvedPublic

Description

See the parent task

Event Timeline

Tip/question: why they were inserted in the firs place? Will we have common duplicate key inserts in the future?

They shouldn't have inserted in the first place. We have two ways to insert the data:

  • Using the jobrunner. I'm guessing that's not the issue.
  • Using the maintenance script when we update the model. Since It might overlap with the job runner.

I'm not sure how we can avoid such cases.

Change 312286 had a related patch set uploaded (by Ladsgroup):
Add maintenance/CleanDuplicateScores.php

https://gerrit.wikimedia.org/r/312286

Change 312286 merged by jenkins-bot:
Add maintenance/CleanDuplicateScores.php

https://gerrit.wikimedia.org/r/312286

Change 314258 had a related patch set uploaded (by Thiemo Mättig (WMDE)):
Revert "Add maintenance/CleanDuplicateScores.php"

https://gerrit.wikimedia.org/r/314258

Change 314258 abandoned by Thiemo Mättig (WMDE):
Revert "Add maintenance/CleanDuplicateScores.php"

Reason:
See Iadf3b00 instead.

https://gerrit.wikimedia.org/r/314258